1. Structural Alterations and Conformational Dynamics in Holliday Junctions Induced by Binding of a Site-Specific Recombinase 
Jehee Lee, Yuri Voziyanov, Shailja Pathania, Makkuni Jayaram 
Molecular Cell 
Volume 1, Issue 4, Pages (March 1998)
DOI: /S (00)

2. Figure 1
Potential Holliday Junction Configurations during Flp Site-Specific Recombination
Three possible conformations of a Holliday junction intermediate formed during Flp site-specific recombination are shown. (A) and (B) represent the right-handed antiparallel stacked X forms, while (C) represents the unstacked square-planar form. The L and R designations refer to the left and right arms of the two DNA substrates, L1R1 and L2R2, from which the Holliday intermediate is formed by exchange between red and orange strands (say, the top strands). In (A), the top strands, which have already been recombined, are continuous; the parental bottom strands (shown in blue and green) are discontinuous (or crossed). The opposite is true in (B). The scissile phosphates at the left and right ends of the spacer are shown by the closed and open circles, respectively. Experiments with the lambda Int protein and the E. coli XerD and XerC proteins have revealed a large cleavage bias toward the crossed strands (Arciszewska et al. 1997; Azaro and Landy 1997). If this rule holds for Flp as well, resolution of the X-form isomer in (A) would yield the recombinants L1R2 and L2R1. On the other hand, resolution of the X-form isomer in (B) would reverse the initial strand exchange to yield the parentals L1R1 and L2R2. It has been proposed that A↔B isomerization may be an integral step in the Int family recombination pathway (Arciszewska et al. 1997; Azaro and Landy 1997). The planar, four-fold symmetric Holliday isomer in (C) could, in principle, be an intermediate during the isomerization step. Note that, in this isomer, all four scissile phosphates are in an equivalent but noncleavable state. Based on the results with Flp, we suggest that cleavage configuration at the left or at the right end of the spacer can be obtained by limited, directional branch migration coupled with modest reconfiguration of the Flp monomers (see Figure 8). This model does not postulate the restacking of helical arms in an X-form Holliday structure. The arrangement of the DNA arms and the recombinase subunits within the recently solved structure of the “cleaved Cre-loxP intermediate” (Guo et al. 1997) would be generally consistent with the Flp model.
Molecular Cell 1998 1, DOI: ( /S (00) )

3. Figure 8
Models for the Flp Recombination Reaction that Utilize Unimodal or Bimodal Cleavage by Flp
In the initial step, (A), the two cleavages (donation of active site tyrosine) are directed from R1 to L1 and R2 to L2, and correspond to the transhorizontal mode (Chen et al. 1992a). The tilt in the DNA arms is intended to schematically represent the Flp-induced DNA bend (Schwartz and Sadowski 1990; Chen et al. 1992b) as well as the selective strand cleavage observed in substrates with strand-specific nucleotide bulges (Lee et al. 1997). In the resultant Holliday junction (B), the branch point is placed 3 nt away from the left cleavage positions (3, 5; Lee and Jayaram 1995; Zhu et al. 1995). This junction is shown to assume a square-planar configuration (C) in accordance with the results of this study. The branch point (4, 4) in this four-fold symmetric state is equidistant from all four cleavage positions. The reaction path diverges at this point between the two models. The path along C→D→E resolves the junction by transhorizontal cleavages (L1 to R1 and L2 to R2), whereas that along C→D′→E′ resolves the junction by transdiagonal cleavages (L1 to R2 and L2 to R1). In both models, the branch point at the time of resolution is shown to be situated 3 nt away from the right end of the spacer (5, 3; D and D′). For details, see text. The mode of “discontinuous” exchange of the spacer DNA between the recombining partners proposed here (in steps of 3, 1, 1, and 3 nt) is compatible with the “strand swapping” model for λ Int recombination arrived at by Nunes-Duby et al Results from XerC/XerD-mediated resolution of Holliday junctions also agree with this model (Arciszewska et al. 1995).
Molecular Cell 1998 1, DOI: ( /S (00) )

4. Figure 2
Conformations of a Holliday Junction Immobilized at the Center of the Spacer in Its Free and Flp-Bound Form
The junction, J1, is schematically represented at the top left with the arms arranged in the conventional parallel orientation. The left arms are designated as L1 and L2, the right arms as R1 and R2. The parallel arrows are the Flp binding elements. The wavy lines represent DNA sequences that are not part of the Flp target. The 5′ end of each strand is indicated by the open circle. The top and bottom strands are drawn in thick and thin lines, respectively. The point of immobilization of the branch point in J1 is between 4 and 4′ positions of the spacer. Positions 0 and 0′ correspond to the bases immediately to the left and to the right of the spacer, respectively. At the top right is a representation of J1 in the antiparallel form, with the branch point assigned its class number and sequence designation according to Table 1 in Altona Note that the junction is named by reading the string of four bases to the 5′ side of the branch point on each strand (shown in bold letters) in a clockwise direction. This nomenclature facilitates comparisons of this junction (as well as others used in this study) with the Int Holliday junctions described in Azaro and Landy The electrophoretic mobilities of the six permuted Holliday species with two long arms and two short arms in the presence of Mg2+ (A) or in the presence of EDTA (B) are displayed at the left. The inferred disposition of the DNA arms are schematically shown at the right. The protein-free junctions were run in the odd-numbered lanes; the Flp(Y343F)-bound junctions were run in the even-numbered lanes. The data from (A) was used to deduce whether the top or the bottom strands assumed the crossed configuration in the stacked X form of the junction. The X form and the square-planar form are diagrammed at the bottom so that the reader can easily see how these structures satisfy the correspondence between the experimentally obtained gel mobilities shown in the left panels and the Holliday species symbolically represented in the right panels. Note that in this figure and the subsequent ones, the junctions were run as follows: L2R1 (lanes 1 and 2); L1R2 (lanes 3 and 4); R1R2 (lanes 5 and 6); L1L2 (lanes 7 and 8); L2R2 (lanes 9 and 10); and L1R1 (lanes 11 and 12). By convention, the name of a junction is denoted by its two long arms.
Molecular Cell 1998 1, DOI: ( /S (00) )

5. Figure 3
Analysis of a Centrally Immobilized Holliday Structure with Purines Flanking the Junction Being Concentrated in the Bottom Strand
The details are as described for Figure 2. Note that in the stacked X form of J2, the bottom strands (thin lines) were in the crossed configuration.
Molecular Cell 1998 1, DOI: ( /S (00) )

6. Figure 4
Analysis of a Centrally Immobilized Holliday Structure with Equimolar Purine-Pyrimidine Distribution of Flanking Bases in the Top and Bottom Strands
Since the six permuted species of J3 in the presence of Mg2+ had virtually identical mobility, it is not possible to determine which strands were continuous, and which ones were crossed. This is indicated by using dashed lines for the strand crossing in the parallel representation of J3 (top left). Similarly, all four strands are shown as being equivalent under the Altona 1996 classification (top right).
Molecular Cell 1998 1, DOI: ( /S (00) )

7. Figure 5
Figure 5. Mobility Patterns of a Holliday Junction with the Branch Point Immobilized near the Left End of the Spacer
The asymmetric location of the crossover point in J4 (between spacer positions 1 and 2) is responsible for the stagger between the bands in lanes 5 and 7 of (A) and (B). Note that only the two species with two short left arms or two short right arms showed the aberrant migration.
Molecular Cell 1998 1, DOI: ( /S (00) )

8. Figure 6
Mobility Patterns of a Holliday Junction with the Branch Point Immobilized near the Right End of the Spacer
As explained under Figure 5, the stagger between the bands in lanes 5 and 7 is due to the asymmetry in the position of the crossover point in J5 (between spacer positions 1′ and 2′). Note that the direction of the stagger for J5 is opposite to that of J4 (immobilized close to the left end of the spacer; Figure 5).
Molecular Cell 1998 1, DOI: ( /S (00) )

9. Figure 7
Probing of Immobilized Junctions with Potassium Permanganate for Unstacked/Unpaired Thymine Residues
Two of the junctions tested, J4′ and J5′, were derivatives of J4 and J5 (see Figure 5 and Figure 6) with the crossover points locked in between spacer positions 1 and 2 or 1′ and 2′, respectively. J6 was a mobile junction with the branch point free to migrate through the spacer as well as the Flp binding elements bordering it. After binding Flp(Y343F), the junctions were subjected to electrophoresis in the presence of EDTA. The bands corresponding to the free and bound junction were excised and treated with KMnO4 in situ in the gel. After extraction from the gel, the DNA was subjected to piperidine cleavage. Lanes 2 and 3 represent the strand cleavage profiles from the permanganate-treated free junction and Flp(Y343F)-associated junction, respectively. Lane 1 represents the free junction subjected to the piperidine reaction without prior KMnO4 treatment. The pattern in lane 1 results primarily from background bands due to some cleavage at the G positions during piperidine treatment. The G+A and C+T Maxam-Gilbert sequence ladder at the left was run for reference. The position of the radioactive label is denoted by the asterisk. The label was placed at 5′ ends in J4′ and J5′, and at 3′ ends in J6. One strand each of J4′ and J5′ and two strands of J6 were probed for permanganate sensitivity. The horizontal arrows indicate the positions of the branch point in J4′ and J5′ in the absence of Flp(Y343F).
Molecular Cell 1998 1, DOI: ( /S (00) )